Journal of Hydrology 598 (2021) 126208

Contents lists available at ScienceDirect

Journal of Hydrology
journal homepage: www.elsevier.com/locate/jhydrol

Research papers

Modeling lake outburst and downstream hazard assessment of the Lower

Barun Glacial Lake, Nepal Himalaya
Ashim Sattar a, Umesh K. Haritashya a, *, Jeffrey S. Kargel b, Gregory J. Leonard c,
Dan H. Shugar d, Donald V. Chase e
a
Department of Geology and Environmental Geosciences, University of Dayton, Dayton, OH 45469, USA
b
Planetary Science Institute, Tucson, AZ 85719, USA
c
Lunar & Planetary Laboratory, Department of Planetary Sciences, University of Arizona, Tucson, AZ 85721, USA
d
Water, Sediment, Hazards, and Earth-surface Dynamics (waterSHED) Lab, Department of Geoscience, University of Calgary, Alberta T2N 1N4, Canada
e
Department of Civil and Environmental Engineering and Engineering Mechanics, University of Dayton, Dayton, OH 45469, USA

A R T I C L E I N F O A B S T R A C T

This manuscript was handled by Marco Borga, Climate change-driven retreat of glaciers is producing thousands of glacial lakes across mountain regions. These
Editor-in-Chief, with the assistance of Marco lakes generally grow, coalesce into larger lakes that may produce increased downstream hazards and risks due to
Toffolon, Associate Editor glacial lake outburst floods (GLOFs). This study assesses such hazards of Lower Barun Lake located near Mount
Everest, Nepal. We model a series of scenarios, including two potential avalanches that enter the lake from the
Keywords:
surrounding slope and eight potential GLOFs from the lake. To evaluate the susceptibility of the frontal moraine
Glacial lake
to overtopping, we characterize the initial avalanche-induced surge of water over the moraine caused by the
GLOF hazard
Himalaya kinetic energy of arriving masses and possible tsunami-like events. Further, we present physical hydrodynamic
HEC-RAS models that reveal the hazard from the potential overtopping and GLOF events along the Barun-Arun river valley.
Special attention is given to analyze the flow hydraulics at six downstream settlements. To estimate potential
impacts at each location, two extreme-magnitude, two high-magnitude, two moderate-magnitude, and two low-
magnitude GLOFs were hydraulically evaluated for the present lake dimension and the modeled future growth of
the lake. As with most hydrological processes, the magnitude and frequency of GLOFs from Lower Barun Lake
have an inverse, albeit uncertain, relationship, but the potential impacts on people and infrastructure are
extremely sensitive to the events’ magnitude. The flow dynamics results indicate that an overtopping flood
without erosion of the damming moraine causes minimal impact in the valley. The extreme-magnitude and high-
magnitude GLOF cases, where the moraine is incised, have a larger impact but differ greatly in magnitude at each
of the downstream settlements. The moderate-magnitude and low-magnitude GLOFs, while the most frequent
type, have limited volume and peak discharge, causing less impact downstream. Our calculations only portray
the part of the hydrograph representing lake overfill due to a volume of ice or rock entering the lake, and the
volume of the lake that could drain from a breach of the damming moraine down to specified depths over
specified time periods.

1. Introduction
Rising global temperatures have increased glacier melting during
this century at rates more rapid than in the previous century (Bolch
et al., 2012; Kraaijenbrink et al., 2017; Roe et al., 2017). As meltwater
becomes more voluminous, runoff can accumulate in depressions on
thinning and sagging glacier surfaces, or in front of receding glaciers,
and form moraine-dammed lakes such as Lower Barun in Nepal (Lins­
bauer et al., 2016; Haeberli et al., 2016; Sattar et al., 2019b). Moraine-
dammed glacial lakes are the most common sources of glacial lake
outburst floods (GLOFs) (Watanbe and Rothacher, 1996; Westoby et al.,
2014). GLOFs have been observed many times since the end of the Little
Ice Age (Vuichard and Zimmermann, 1987; Kargel et al., 2011; Carey
et al., 2012; Clague and O’Connor, 2015; Vilímek et al., 2015; Anacona
et al., 2015a; Paul and Bolch, 2019). However, GLOFs emanating from
moraine-dammed lakes worldwide became far more common in the
1930s, an increase believed to be due to lagging responses to the end of
the Little Ice Age. Since then, the incidence of GLOFs has fluctuated, first

* Corresponding author.
E-mail address: uharitashya1@udayton.edu (U.K. Haritashya).

https://doi.org/10.1016/j.jhydrol.2021.126208
Received 21 September 2020; Received in revised form 25 January 2021; Accepted 10 March 2021
Available online 16 March 2021
0022-1694/© 2021 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
A. Sattar et al.
rising sharply and then falling somewhat since the 1980s (Harrison
et al., 2018); yet it is expected to increase within parts of the Himalaya in
the coming decades (Veh et al., 2020).
Moraine-dammed lakes are confined by end moraines or lateral
moraines, with ice forming the up-valley lake margin until the glacier
retreats farther and detaches. These moraines, however, usually consist
of unconsolidated sediment, thereby making dam structures unstable
compared with bedrock dams. Many of the moraines, especially in the
Himalaya, initially contain ice (Richardson and Reynolds; 2000), which
provides greater structural strength but eventually melts, leaving un­
consolidated sediment and poor structural support to the dam. As the
dams degrade due to erosion or melting of the ice, the lakes may grad­
ually lower or become infilled with sediment, and slowly disappear.
Sudden GLOFs from moraine-dammed or any other types of naturally
dammed glacial lakes (supraglacial or ice-marginal) sometimes occur
and pose risks to downstream infrastructure and communities (Osti and
Egashira, 2009; Kropáček et al., 2015; Allen et al., 2016). As lakes
continue to grow in number and size globally (Shugar et al. 2020), so
does the cause for concern. GLOFs from moraine-dammed lakes may be
triggered by (1) overtopping and erosion of the moraine dam caused by
waves generated from ice avalanches, rockfalls, or calving; (2) meteo­
rological conditions, especially extreme rain events, which may partly
melt ice in the moraine dam and weaken the moraine, or cause overfill
and thermal and physical erosion of the moraine, or induce a mass
movement into the lake; (3) degradation from within the moraine dam
(melt-through due to seepage from the lake and physical erosion—pip­
ing; or flotation of the moraine’s buried ice by the hydrostatic gradient
imposed by the lake); (4) a small GLOF, perhaps from a supraglacial
pond, may overfill a moraine-dammed lake; or (5) seismic activity, when
strong ground motion destabilizes the damming structure (Richardson
and Reynolds, 2000; Westoby et al., 2014). These potential triggers each
involves a stochastic chain reaction sequence, which cannot be explicitly
predicted but may be understood either statistically for a population of
similar lakes or individually through scenario modeling. Here, because
we are investigating one lake, we take the latter approach.
All trigger mechanisms associated with GLOFs can range in magni­
tude and severity, depending on the lake’s and terrain’s individual de­
tails, including: (1) the lake’s total drainable volume; (2) the internal
sedimentological structure of the moraine dam, which can have low
internal friction and high propensity to break up suddenly and cause
high, peaked outbursts; or alternatively can have high internal friction
and produce robust boulder armoring of drainage conduits and thus
limit erosion and reduce peak discharge; (3) the amount, distribution,
and structure of buried ice in the end moraine, which can control the
likelihood and rate of melt-through piping, and can cause large changes
in the moraine’s structural integrity; (4) the width-to-height ratio of the
moraine dam, for instance, a dam that is sharply peaked can readily and
quickly break up, whereas a broad dam is more robust; (5) the height of
the moraine dam, both the freeboard, which controls the damage and
amount of overtopping for a given size of the wave, and the down-valley
slope that partly controls the hydraulic pressure; (6) the lake’s position
with respect to its surrounding topography, which can either allow large
mass movements to enter the lake rapidly and with large and sudden
kinetic energy input, or not; (7) the mass of unstable hanging glaciers/
snow or perched, fractured bedrock on adjacent mountain slopes; or (8)
the lake bed and shore structure that can either cause tsunami waves to
ramp up sharply, or undergo constructive wave interferences, or cause
tsunami waves to diffuse harmlessly.
We note that one recent example of a mass movement-induced
displacement wave caused overtopping of a moraine dam having 80 m
of freeboard, so these waves can be rather extreme (Hubbard et al.,
2005; Emmer, 2017). More modest impulse wave overtopping by about
5 m, caused, for example, by a large ice avalanche into Lake 513 in Peru
(Worni et al., 2014), and similar to the smallest avalanche-triggered
events modeled in this paper, can nonetheless generate damaging
debris flows.
2
Journal of Hydrology 598 (2021) 126208
The rapid growth in the size and number of glacial lakes, along with
rapid increases in infrastructure development (e.g., hydropower plants,
roads, and bridges) and population in mountainous regions have cata­
lyzed the need for: comprehensive regional and global mapping of
glacial lakes (e.g., Campbell and Pradesh, 2005; Bajracharya and
Shrestha, 2011; Emmer et al., 2016; Shukla et al., 2018; Shugar et al.,
2020), comparative hazard and risk assessment (Thakur et al., 2016;
Sattar et al., 2019a; Maskey et al., 2020), GLOF reconstruction (Majeed
et al., 2020), prioritization for mitigation efforts (Khanal et al., 2015),
assessment of earthquake aggravation of GLOF hazards (Kargel et al.,
2016), detailed field, remote sensing, and modeling evaluation for GLOF
hazard mitigation (Somos-Valenzuela et al., 2015; Kargel et al., 2016),
and implementation of mitigation designs and consideration of coupled
mitigation and economic development opportunities at glacial lakes
(Regmi et al, 2017; Emmer, 2018).
In disaster risk assessment and management, risk conventionally
involves the confluence of hazard, vulnerability, and exposure (e.g.,
Lavell et al., 2012); or if these are quantified, then the risk is the product
of these components. The relative risk a GLOF may present to the
downstream region depends on the hazard components consisting of
peak discharge and cumulative discharge, which are functions of the
drainable volume of the lake, condition of the damming moraine, po­
tential GLOF triggers (listed above), and overtopping due to external
impact which together determines the hazard; and the exposure and
vulnerability of elements downstream of the lake (infrastructure, human
life, and ecological values, including the resilience of these at-risk
components). The risk that a potentially dangerous lake may present
to the low-lying areas may be unrecognized unless a detailed hazard and
risk assessment are undertaken. GLOF hazard assessments often contain
inherent uncertainties as outburst events are hypothesized to occur in
the future due to a variety of potential triggering processes. However,
applying variable assumptions and conditions within hydrodynamic
modeling, based on the spatial and geological characteristics of a lake
and its surroundings, allows us to understand how a valley will behave
in a lake failure event. Furthermore, developing detailed assessments
and models addressing extreme-magnitude scenarios are needed for
responsive disaster preparedness and mitigation.
The purpose of this study is to evaluate the GLOF hazard associated
with Lower Barun Lake, located in the Nepal Himalaya, by performing a
series of modeling and downstream flood routing. Given the complexity
of the possible GLOF triggers in the mountainous basin, our first
objective includes identifying and modeling potential avalanche around
the lake to evaluate the susceptibility of the damming moraine to
overtopping and downstream propagation of the overtopping flow.
Although we do not model the erosion of the frontal moraine, some
points about this erosion are considered in Supplementary section-1.
The second objective is focused on GLOF hazard assessment by assuming
that an overtopping wave can initiate erosion irrespective of the trigger
and lead to complete breaching of the moraine. Thus this objective can
be broadly categorized into two parts (i) potential GLOFs originating
from Lower Barun Lake at its current state (present lake), under the
following scenarios: extreme-magnitude, high-magnitude, moderate-
magnitude, and low-magnitude GLOFs (ii) consider the same scenarios
for the maximum future extent and estimated volume of the lake. The
latter builds on the first but considers the future volume and extent of
the lake for all four scenarios. The magnitude in each scenario relates to
modeled flood volume, which integrates the breach parameters, as flood
volume tracks with both breach values on Table 1. To achieve the first
objective, we used remotely sensed data, field-collected information of
potential avalanches that could trigger a GLOF, and two-dimensional
dynamic physical models, including RAMMS (Rapid Mass Movement
Simulation) and HEC-RAS. For the second objective, we used remotely
sensed data, field-collected lake bathymetry (Haritashya et al., 2018)
and HEC-RAS hydrodynamic modeling. We do not consider the in­
stances of cascading hazards (e.g., Kirschbaum et al. 2019), such as
GLOF spilling from nearby Upper Barun Lake into Lower Barun Lake,
A. Sattar et al. Journal of Hydrology 598 (2021) 126208

Table 1 impact of the different magnitude of floods originating from the lake in
Summary of the different GLOF scenarios. Future GLOF is based on a future lake the present and the future, irrespective of the trigger mechanism.
extent.
Scenarios GLOF Description PFV/Vw hb Tf Bw 2. Study area background and importance
(*106m3) (m) (h) (m)

1 Extreme-magnitude 101.2 104 0.6 222.1 GLOF studies in the Hindu Kush-Karakoram-Himalaya (HKKH) pre­
present-lake viously ranged from regional-scale assessments (Quincey et al., 2007;
2 High-magnitude 78.8 64 0.9 187.1 Worni et al., 2013; Nie et al., 2017; Veh et al., 2018) to site-specific
present-lake
3 Moderate-magnitude 46.5 32 1.3 138.5
hazard assessments (Osti and Egashira, 2009; Bajracharya and Mool,
present-lake 2009; Fujita et al., 2009; Shrestha et al., 2010; Sattar et al., 2019a,
4 Low-magnitude present- 16.0 16 1.4 86.4 2019c). In the Nepal Himalaya, glacial lake studies include compre­
lake hensive glacial lake inventories (Reynolds and Taylor, 2004; Bajra­
5 Extreme-magnitude 179.0 104 0.9 266.8
charya and Mool, 2009; Khadka et al., 2018), identification of hazardous
future-lake
6 High-magnitude future- 139.5 66 1.2 225.9 glacial lakes (Rounce et al., 2017), GLOF return period (Cook et al.,
lake 2018), and GLOF reconstruction (Byers et al., 2018). Among the largest
7 Moderate-magnitude 82.8 33 1.7 167.6 glacial lakes in Nepal, Imja Lake (Imja Tsho) has been the most widely
future-lake studied, including detailed GLOF modeling (Somos-Valenzuela et al.,
8 Low-magnitude future- 26.9 16.5 1.8 102.5
lake
2015; Lala et al., 2018), field-based geophysical investigation of the
frontal moraine (Dahal et al., 2018), investigations of its evolution over
the past few decades (e.g., Fujita et al., 2009; Haritashya et al., 2018),
because our focus is on the direct inputs of solid mass into Lower Barun and flood risk and management (Cuellar and McKinney, 2017).
Lake and downstream GLOF hazard assessment. The series of outburst In the last two decades, however, nearby Lower Barun glacial lake
flood events modeled in this study are used to evaluate the downstream (27◦ 47′ 51′′ N, 87◦ 05′ 26′′ E) in the Makalu region of Nepal has been

Fig. 1. Overview of the study area showing Lower Barun Glacier and Lower Barun Lake located in the Barun Khola basin, Nepal. The settlements (red stars) and
hydropower sites (cyan triangles) are marked along the flow channel. The distances of the selected sites from the lake are 29.7 km (Barun Bazaar), 32.4 km (Site-1),
38.9 km (Site-2), 40 km (Site-3), 49 km (Site-4), and 70 km (Site-5).

3
A. Sattar et al. Journal of Hydrology 598 (2021) 126208

growing rapidly, nearly doubling in area since the 1960s (to 2.2 km2 in
2019) (Fig. S4). The lake is oriented east–west within the upper Barun
drainage basin (Fig. 1) and it sits at an elevation of about 4550 m. The
Barun River — a tributary of the Arun River (Sankhuwasabha District,
Eastern Nepal) — drains from Lower Barun Lake and Barun Glacier,
which feeds the lake. The Barun River also derives flow from Upper
Barun Lake (and other smaller lakes in the basin), which drains down
alongside Lower Barun Lake but is separated by a lateral moraine. As
Upper Barun Lake itself poses a significant GLOF hazard, small future
changes in drainage could produce a formidable cascading hazard sit­
uation. For now we only evaluate Lower Barun’s hazard without
involvement of Upper Barun Lake.
Lower Barun Glacier is characterized by numerous supraglacial
lakes, ice cliffs, and thick debris cover in the lower ablation zone (Fig. 2
a-b). Lower Barun Lake is dammed by an ice-cored moraine (Dahal et al.,
2018; Haritashya et al., 2018) on the east and the Barun glacier’s ter­
minus to the west. The damming moraine has a set of ponds through
which the lake drains into and feeds the Barun River. In this regard, it is
similar to Nepal’s Imja Lake, and unlike Thulagi and Rolpa lakes, both of
which drain directly over narrow end moraines without intervening
ponds. Direct drainage over narrow, steep end moraines makes Thulagi
and Rolpa lakes more hazardous or more likely to drain in a shorter
period of time than Lower Barun and Imja lakes. However, of all these
lakes, Lower Barun is by far the deepest known lake with the largest
drainable volume, is the fastest-growing, and has the most rapid accel­
eration of growth (Haritashya et al., 2018; see Fig. S4); and the adjacent

Fig. 2. (a) Field evidence of the retreating and thinning glacier terminus; (b) steep ice cliff upstream of the present glacier terminus; (c) bathymetric survey on a
kayak mounted with SyQwest Hydrobox Sonar sensor; (d) Lower Barun glacier terminus.; (e) the Lower Barun glacier and lake showing the high and steep slopes (45-
55 degrees) at locations of potential avalanches that could enter the lake.

4
A. Sattar et al.
mountain relief is immense. Furthermore, Lower Barun’s end moraine is
undergoing thermokarstic evolution, as is Imja Lake’s, such that one of
two things will happen: either the melting of ice within the end moraine
will gradually lower the lake, or else the end moraine will evolve to a
structure like that of Rolpa and Thulagi lakes, with a precarious, narrow
end moraine dam. While the final state of Lower Barun’s end moraine
cannot be predicted with certainty, the hazard associated with moraine
failure is increased if the moraine evolves into a narrow structure.
Hazards are reduced if interior moraine-ice melts and lake levels fall.
The Barun valley has witnessed at least two known GLOF events
(Meon and Schwarz, 1993; USAID, 2014; Byers et al., 2018). The first
event flooded the Arun River in 1964 and was witnessed by the local
people, but no other information about the source of the flood exists
(Yamada and Sharma, 1993). The second outburst event occurred on
April 17, 2017, and involved a rockfall and avalanche into Langmale
glacial lake, located ~3 km east of Lower Barun Lake (Byers et al.,
2018). The resulting flood deposited a huge volume of debris along the
Barun River flow channel and impacted the settlements of Langmale and
Yangla Kharke (6.5 km downstream, Fig. 1), containing Buddhist gom­
pas and a few teahouses. This event blocked the confluence of Arun and
the Barun rivers (29 km downstream from Lower Barun lake) and
formed a short-lived landslide dammed lake. Yamada and Sharma
(1993) identified Lower Barun Lake as potentially dangerous based upon
the examination of aerial photographs. However, Kattelmann (2003)
deemed Lower Barun Lake to be less threatening than other lakes in the
broader region, such as Imja Lake, primarily due to the sparsely popu­
lated downstream regions. In any case, Lower Barun Lake has grown
considerably since 2003 (Haritashya et al., 2018) exposing it to large
mountain sides, and the Langmale event indicate that some rock units
and mountain peaks in the valley exhibit large masses of unstable rock
(Byers et al., 2018), posing a far greater threat than nearly two decades
ago.
Of particular concern today are potential impacts of outburst floods
on existing and proposed regional hydroelectric projects downstream of
Lower Barun Lake. For example, the powerhouse of the proposed $1.4B,
1061 megawatt (MW) Upper Arun Hydroelectric project is located 2 km
downstream from the confluence of Barun and Arun rivers (immediately
above Site-1 in Fig. 1) (https://kathmandupost.com/money). Moreover,
the 900 MW Arun III Hydropower project, also on the Arun River, has its
starting tunnel located 49 km downstream of Lower Barun Lake
(immediately west of Site-4 in Fig. 1). Both of these projects would be
completely exposed to a GLOF from Lower Barun Lake. Therefore, it is
important to investigate the range of potential GLOF hazards as a pre­
cursor to the kind of detailed physical field-science studies that preceded
the engineered partial mitigation (lake lowering and stabilization of
drainage outlet) of Imja Lake’s GLOF hazard in 2014 (UNDP Imja
Report, 2013; http://cfgorrp.dhm.gov.np/). Recently, several small
settlements have formed along the lower Barun and Arun valleys
(identified using multi-temporal high-resolution Google Earth imagery),
and new hydropower projects could further spur growth in this valley,
potentially exposing them to large GLOF risks in the future. Our study
includes the potential impact of various GLOF scenarios on several
downstream sites (Barun Bazaar and Site-1 to Site-5 as shown in Fig. 1).
Apart from these settlements, numerous bridges and other small infra­
structure exist along the channel downstream from Lower Barun Lake to
Site-5.
3. Data
Bathymetric data were acquired during field investigation in October
2015 using an inflatable kayak-mounted SyQwest Hydrobox sonar echo-
sounding sensor and modeled within ArcGIS using natural neighbor
interpolation. More details about the methodology related to bathy­
metric observations, modeling, and uncertainty can be found in Har­
itashya et al. (2018). Lake bathymetric maps permitted an accurate
simulation of the time-varying drawdown of Lower Barun Lake.
5
Journal of Hydrology 598 (2021) 126208
Bathymetric information was also necessary to develop relatively ac­
curate moraine breach models and determine the potential flood volume
(PFV) released in different breach scenarios.
In avalanche simulations and hydrodynamic modeling, terrain
properties such as surface elevation and surface roughness are often
extracted using digital elevation model (DEM) and Landuse and Land
Cover (LULC) composites, respectively. Here we used an Advanced Land
Observing Satellite (ALOS) – Phased Array type L-band Synthetic
Aperture Radar (PALSAR) DEM with a spatial resolution of 12.5 m. The
ALOS PALSAR DEM is a radiometrically terrain-corrected elevation
product released globally in October 2014 by the Alaska Satellite Fa­
cility (https://asf.alaska.edu/data-sets/derived-data-sets/alos-palsar-rt
c/alos-palsar-radiometric-terrain-correction/). It has been successfully
used for dynamic GLOF routing previously in the Himalaya (Dhote et al.,
2019; Sattar et al., 2019a; Maskey et al., 2020). Most models either use a
constant value of Manning’s roughness coefficient or separate values for
upstream and downstream routing. We used Landsat-8 derived LULC
product (30 m) (Uddin et al., 2015) to extract pixel-based Manning’s
roughness coefficients along the flow channel. This enables us to define
ground-channel friction in a spatially distributed pattern for accurate
hydraulics.
For glacier bed mapping and future lake-volume estimation, we used
the ALOS-PALSAR DEM, glacier outlines (RGI 6.0), and manually digi­
tized glacier flowlines as inputs to generate an ice-thickness model
(Linsbauer et al., 2012). We also employed the ensemble ice-thickness
(Farinotti et al., 2019) and DEM for a comparative analysis of the
glacier bed. Here, the glacier ice-thickness distribution (RGI160-
15.03366) (Farinotti et al., 2019) was used to derive the glacier bed and
the frontal overdeepening extent.
4. Methodology
The entire methodology of this study is summarized in Fig. 3, and
details of each component are presented below.
4.1. Future lake extent and volume
To calculate the lake’s future volume, we estimate the maximum
extent of Lower Barun Lake by mapping the glacier bed using a spatially
distributed glacier ice thickness product and DEM (Fig. 3 and Fig. 5a).
Lacking ice thickness soundings and subglacial bed topography, we
instead used the Glacier bed Topography (GlabTop) method (Linsbauer
et al., 2012) and the ensembled ice thickness (Farinotti et al., 2019; see
section 3 for data used) to calculate the glacier bed of Lower Barun
Glacier. GlabTop is a shear-stress based model used to calculate ice
thickness (h) for alpine glaciers, described by h = τ /(f ρg sin α), where f
is the shape factor, ρ is the ice density, g is the gravitational acceleration,
and sin α is the slope of the glacier. The ensembled ice-thickness is an
average product derived using four different ice thickness models
(Farinotti et al., 2019). We used three different shape-factors for glacier
bed calculation using GlabTop (f = 0.6, 0.7, and 0.8). We also completed
a raster-based calculation of ice thickness to derive the distributed
glacier bed for all three values of f. An overdeepening site is extracted
using raster-based operations in a GIS platform (ArcMap 10.7.1). By
“overdeepened,” we mean that the glacier bed extends to a lower
elevation than the areas immediately downvalley from the lake.
The future lake extent and volume are obtained by interpolating
between the current lake bathymetry (resampled to 25 m) and the
modeled overdeepened bathymetry of the glacier-bed (e.g., Fig. 5b). We
compare the four different modeled glacier-bed outputs (GlabTop-0.6,
0.7, 0.8, and ensemble output) with the glacial lake extent and field
measured depths. The validated glacier bed is further used to calculate
the volume of the frontal overdeepening site of the glacier. The future
lake volume for our hazard assessment analysis is calculated by
combining the current lake volume (from bathymetry data) and the
volume of frontal overdeepening (both spatially resolved at 25 m). This
 A. Sattar et al.
6

Journal of Hydrology 598 (2021) 126208

Fig. 3. The overall methodology adopted in the present study.
A. Sattar et al.
method to identify and map glacier bed overdeepening has been suc­
cessfully used both at the global scale as well as locally in the Himalaya
(Linsbauer et al., 2016; Frey et al., 2010; Haeberli and Linsbauer, 2013;
Sattar et al., 2019b). The growth model also provides a good indication
of how closely the west end of the lake will approach the high and steep
adjacent mountain relief shown in Fig. 2e.
4.2. Potential GLOF triggers and overtopping
We identify two potential weak zones around the lake from where ice
avalanches could trigger any of the GLOFs considered in this study
(Fig. 2e); hence, we have modeled the avalanche impacts to determine if
their impact waves overtop the frontal moraine. The avalanche locations
were not picked arbitrarily but are based on our field observation
(Fig. S5) and hazard assessment of Rounce et al. (2016) that identified
slopes between 45◦ and 60◦ as being avalanche-prone, and where glacier
ice masses appear to hang precariously. Most of these slopes host snow/
ice perched atop a steep bed, which can be a potential trigger for any of
the modeled GLOFs in this study. A significant avalanche emanated from
one of these slopes during fieldwork, and a time-series of the avalanche
has been captured (Fig. S5). The average speed of the avalanche as it
crossed the lake was ~100 km h− 1 (27.7 m s− 1), calculated from the
time it took to cross the width of the lake. That is both measured and
partly inferred by knowing approximately when it first reached the lake
as the camera location and the route down the mountain is known, and
the vectors from the camera to the lake shore can be closely estimated.
The avalanche did not lead to the overtopping of the lake. Therefore, we
model a higher magnitude potential-avalanche originating from the
same location (Avalanche-1). For the second trigger (Avalanche-2) that
can potentially have an impact on the future lake, we recognize another
major slope with hanging snow/ice that occurs adjacent to the site
described above, as a potential avalanche source zone. As discussed
above and highlighted by Haritashya et al. (2018), the lake could grow
up to this point towards its western end, rendering it vulnerable from
large-scale avalanching from the surrounding slopes. Another important
potential GLOF trigger is Upper Barun Lake, which could drain suddenly
into Lower Barun Lake. However, that is a unique trigger mechanism
that we are not evaluating in this paper.
The severity of these ice/snow mass movement triggering events has
a relation to the volume of water displaced and the condition of the
moraine. Thus, impulsive expulsion of water by the kinetic energy of a
mass input is considered. We use RAMMS to model two potential ava­
lanches. RAMMS is based on Voellmy-Salm finite volume method that
solves the mass flow equations in two-dimension (Christen et al., 2010).
The flow rheology of a mass movement is governed by the release area,
volume, density of the material, friction parameters, and the underlying
topography. Here, we define the release area for two potential ava­
lanches by selecting the contiguous pixels with high slope angles
(45◦ –60◦ ). We define ice/snow thickness as 4.7 m and 4.9 m for
Avalanche-1 and Avalanche-2, respectively, releasing a total volume of
11.2 × 105 m3 of ice/snow (Rounce et al., 2016) in each case. The
density (ρ) of flow material is considered as 1000 kg m− 3 (ice with some
rock debris and liquid water). Friction parameters include the Coulomb
friction (μ) and turbulent friction (ξ) (Bartelt et al., 2013), which were
set as 0.12 and 1000 m s− 2 for μ and ξ, respectively. These input values
are widely used to model avalanches in snow-covered and glaciated
terrain (Schneider et al., 2014; Somos-Valenzuela et al., 2016; Frey
et al., 2018). The outputs were obtained as avalanche runout distances
(m), flow depths (m), and flow velocity (m s− 1).
The avalanche-induced impact waves generated by the potential
avalanches were modeled empirically using the Evers et al. (2019)
model. A previous approach of Heller and Hager (2010) has been used
for avalanche-induced impulse-wave characterization in glacial lakes in
the Himalaya (Lala et al., 2018; Byers et al., 2018, 2020). Based on the
same principles as Heller and Hager (2010), the approach of Evers et al.
(2019) computes the overtopping wave height, overtopping velocity,
7
Journal of Hydrology 598 (2021) 126208
and volume due to impulse waves based on updated studies. The inputs
to the model, such as impact velocity, impact volume, slide thickness,
and slide width, are derived from the RAMMS outputs. We calculated the
impact volume as the volume entering the lake at the end of an
avalanche. The impact-wave run-up angle and freeboard were calcu­
lated using DEM and high-resolution Planet data. Based on the over­
topping characteristics of the impact wave, the overtopping volume (in
m3) and duration (in sec) from Lower Barun Lake is determined for each
impact (Avalanche-1 and Avalanche-2). Further, the overtopping
hydrographs were calculated based on the total overtopping volume and
duration were used as an input for downstream flood routing. We use
HEC-RAS (v 5.0.7) (https://www.hec.usace.army.mil/software-/hec-
ras/), a two-dimensional hydrodynamic model to evaluate the down­
stream propagation of the overtopping wave. The hydrodynamic mod­
ule solves the depth-averaged two-dimensional shallow water equations
(Saint-Venant Equations) in an unsteady flow. The time-series of flow
depth (Df) and flow velocity (Vf) obtained as output in a spatially
distributed manner are used to evaluate the GLOF hydraulics down­
stream of the lake. The detailed model settings for 2D flood routing are
given in the Supplementary section-2.
4.3. Present and future GLOF hazard assessment and scenario modeling
Since overtopping flows can induce dam erosion, forming an initial
breach followed by progressive breach enlargement (Singh, 1996), we
assume the same in the case of Lower Barun Lake. Therefore, over­
topping modeling is followed by dam-breach flood (GLOF) modeling.
Here we evaluate a series of potential GLOFs that can originate from
Lower Barun Lake assuming moraine breaching of different magnitudes,
potentially triggered by the overtopping wave, and some scenarios
implying that a dam breaching erosion event has occurred. Whereas we
include a plausible functional form of the developing breach (a sine
function), the function we select is not an actual erosional model.
Therefore, our GLOF modeling is based on the hypothetical magnitude
of the water released (PFV) and the release duration, not the sedimen­
tological erosion modeling associated with the overtopping.
We use HEC-RAS (v 5.0.7) one-dimensional and two-dimensional
hydrodynamic model to study eight potential GLOF events based on
the varied breach parameters that represent current lake conditions
(scenarios 1–4) and maximum future lake extent (scenarios 5–8) while
keeping moraine dam geometry, downstream valley geomorphology,
and river hydrological conditions the same in every simulation. (Table 1;
Fig. 4). HEC-RAS has been used successfully in one and two-dimensional
hydrodynamic modeling of outburst floods in steep terrain like the
Himalaya (Klimeš et al., 2014; Anacona et al., 2015b; Wang et al., 2018;
Sattar et al., 2019a, 2019c, 2020). Our modeling assumes the continu­
ation of the lake’s rapid observed growth, but by including a wide range
of potential GLOF scenarios, the effects of changing moraine-breach
properties are indirectly included.
Each scenario has a unique set of parameters, including the height of
the breach (hb), time of moraine failure (Tf), breach width (Bw); all of
which constrain the flood volume (Vw) released in a potential outburst
event and are calculated using Froehlich (1995b):
Bw = 0.1803Ko (Vw )0.32 (hb )0.19 (1)
Tf = 0.00254 (Vw )0.53 (hb )− 0.9
(2)
The high accuracy and low prediction error of these sets of equations
make it the most widely used empirical approach for modeling earthen-
dam failures (Wahl, 2004). All the scenarios were modeled as a sine
wave progression breach model where the initial breach forms slowly
and speeds up with time as outflow velocities and shear stress increases
through the breach. The output is obtained in the form of a breach
hydrograph with a peak discharge (Qmax) for each modeled event
considered for downstream routing along the flow channel. Each output
A. Sattar et al. Journal of Hydrology 598 (2021) 126208

Fig. 4. Schematic showing the different dam breach scenarios of Lower Barun Lake, where Scenario-1 and Scenario-5 represent the potential extreme-magnitude
events, Scenario-2 and Scenario-6 represent high-magnitude events, Scenario-3 and Scenario-7 are moderate-magnitude events, and Scenario-4 and Scenario-8
are low-magnitude events; the solid red lines show the respective breach heights. Note that the scenarios are modeled with or without any specific trigger.

hydrograph was evaluated by comparing it with empirically calculated 4.4. Dynamic routing and hydraulic assessment along the flow channel
peak outflow-discharge (Froehlich 1995a):
Two-dimensional hydrodynamic routing of the breach hydrographs
Qp = 0.607Vw0.295 h1.24 (3)
w for each breach scenario (scenarios 1–8) was performed using HEC-RAS
along the flow channel from Lower Barun Lake to Site-5, located 70 km
here, we set up dam-break models for two extreme-magnitude events
downstream. The hydraulic behavior of the GLOF wave was analyzed at
(scenario-1 and scenario-5), two high-magnitude events (scenario-2 and
six different sites (Barun Bazaar, and Site-1 to Site-5; Fig. 1) along the
scenario-6), two moderate-magnitude events (scenario-3 and scenario-
flow channel based on flood discharge (Q), flow depths (Df), and flow
7), and two low-magnitude events (scenario-4 and scenario-8), all
velocities (Vf). The Manning’s roughness coefficient associated with the
originating from Lower Barun Lake.
channel is extracted for the flow area using a Landsat-8 derived LULC
Our field investigation and analysis of remotely sensed data suggest
(30 m) of Nepal (Uddin et al., 2015). Our pixel-based Manning’s
that the frontal moraine has a hummocky surface signifying the presence
roughness distribution (n) for the given flow regime ranged from 0.035
of buried ice. Since no geophysical investigation of the frontal moraine is
to 0.17 based on the different LULC classes present along the flow area.
available, it is unclear how deep this buried ice is. Supplementary sec­
Our approach using Landsat imagery can capture the large-scale
tion-1 shows that in one sector, the debris cover is very thin, but it
roughness but not 1- to 30-m-scale roughness elements, so it is merely
certainly is not thin everywhere. As a result and relying on our field
an approximation to what might be encountered by an outburst. It may
knowledge of the lake’s frontal moraine, the breach height in an
apply strictly only to the largest modeled outburst events but could also
extreme-magnitude case is considered till the point where the hum­
apply to the smaller events if there is a scale-independent (fractal)
mocky terrain ends, where the height appears to be 104 m. This moraine
roughness over the range of scales of the modeled floods. The detailed
height is consistent with the findings of Reynolds and Richardson (2000)
model settings are given in the Supplementary section-2. The model
who found that the terminal moraines from the Neoglacial maxima often
outputs were analyzed to evaluate the flow hydraulics of the GLOF wave
exceed 100 m in height where the resulting lakes can contain 1–20
along the channel. The result metrics are represented in the form of flood
million m3 of water, presenting a serious hazard threat to the down­
inundation limits (If) (km2), discharge (Q) (m3 s− 1), flow velocity (Vf) (m
stream regions. Furthermore, the more widely studied Thulagi lake,
s− 1), and depth (Df) (m) at which the evacuated lake water travels down
located in the Upper Marsyangdi catchment, northern Nepal, is dammed
the Barun Valley.
by a debris-covered stagnant ice body 100 m thick that dates back to the
last glacial advance (Hanisch et al., 1998, 1999; Pant and Reynolds,
5. Results
1999). Therefore, for modeling the extreme-magnitude GLOFs of Lower
Barun Lake, it is assumed that breach incision occurs until it reaches the
5.1. Future lake extent and volume
base of the moraine (i.e., hb = 104 m; Table 1). Further, in the case of
high-magnitude events, hb is considered to be the mean depth of the lake
The Lower Barun future lake extent and volume were calculated
(Fujita et al., 2013) and the corresponding Vw is calculated based on the
using an ice-thickness approach where the present bathymetry of the
lake bathymetry. In the moderate-magnitude scenarios, hb is assumed to
lake was combined to the frontal overdeepening extent of the Lower
be half of the high-magnitude breach depths. And, in the low-magnitude
Barun glacier. Of the four different outputs of modeled overdeepening
GLOFs, hb is half of the moderate-magnitude breach. The PFV from the
(GlapTop f = 0.6, f = 0.7, f = 0.8, and ensembled output), the ensemble-
lake in any of the above GLOFs depends on the bed topography (ba­
model results (Farinotti et al., 2019) correlated most accurately to the
thymetry) and breach dimensions of the frontal moraine. We calculate
present lake extent with a difference of <5% (Fig. S6). The ensembled
the drainable flood volume based on the height of the breach (hb), as the
ice thickness of Lower Barun Glacier has a maximum thickness of 258 m
present bed topography of Lower Barun Lake does not allow its entire
with surface elevation ranging between 4551 and 6669 m a.s.l. The
volume to drain, even in an extreme-magnitude breach (i.e., volume
depths of the frontal overdeepening range up to ~197 m (Fig. 5b). The
below 104 m hb does not drain) (Fig. 6a). The assumed floods are of
spatially distributed bathymetry of the overdeepening site calculated
water only and do not include ingested sediment in the total emitted
using ice thickness and surface elevation data (Farinotti et al., 2019)
flood volumes, the dynamical behavior of the flood (such as impulsive
occupies a total area of 1.62 km2 extending approximately 2 km up­
ejection of water from below hb), or downstream deposited sediment.
stream from the present glacier terminus (Fig. 5b and Fig. 6b, c). The
The hydraulic flood modeling also does not include bank erosional or
overdeepening zone terminates where the bed elevation abruptly rises
depositional changes in channel/valley geometry during the events.
into an ice cliff (Fig. 5c and Fig. 6b). The field-based lake bathymetry

8
 A. Sattar et al.
9

Journal of Hydrology 598 (2021) 126208

Fig. 5. Graphical representation of the raster-based operations for (a) mapping the glacier bed using spatially distributed ice thickness and DEM (Farinotti et al., 2019); (b) resampling and mosaicking depth distributions
of the frontal overdeepening and the lake bathymetry to calculate (c) the future depth/volume of Lower Barun Lake.
A. Sattar et al. Journal of Hydrology 598 (2021) 126208

Fig. 6. (a) Cross-sectional profile of Lower Barun glacier and lake structure; (b) Glacier bed of the Lower Barun glacier; (c) 3-D (East-view) bathymetry of the future
lake showing the frontal overdeepening, present lake extent, and the extreme-magnitude breach level.

reveals a maximum depth of 205 m with the deepest point located near
the current terminus (Fig. 5b). The deepest point of the overdeepening
(ensemble) occurs at about the same location as that of the lake ba­
thymetry, i.e., if the ice were removed, the new lake bed would be about
197 m deep. The depth at the deepest point of the overdeepening (197
m) has a difference of < 5% compared to the bathymetric lake depth at
that given point. The average difference in the depths (overdeepening
bathymetry minus lake bathymetry) of the overlapping area (Fig. 5b and
S6) is calculated to be less than 3%. The maximum depth of the future
lake derived after combining the spatially distributed current lake ba­
thymetry and overdeepening bathymetry is 205 m, which shows that the
lake has already reached its deepest point. The future volume of Lower
Barun Lake calculated by combining the bathymetries of the modeled
frontal overdeepening zone plus the field-based lake bathymetry is
193.5 × 106 m3, covering a total area of 2.69 km2. This volume was used
for future-lake GLOF hazard modeling, as discussed in the next section.

Fig. 7. (a) Flow depth (m), and (b) Flow velocity (m s− 1) of the two modeled potential avalanches; (c) modeled impulse waves and overtopping (Evers et al., 2019);
(d) photograph showing the steep slope (avalanche source zone) located at the right surrounding slope, the modeled avalanche trajectory, and the impact site on the
lake; (e) photograph of avalanche from the right slope witnessed during fieldwork.

10
A. Sattar et al.
5.2. Avalanche modeling and overtopping
We modeled two potential avalanches with their sources from the
surrounding south slope of the lake, one of which has witnessed a pre­
vious avalanche recorded by the field team (Avalanche-1; Fig. S5 and
Fig. 7e). In Avalanche-1, a total of 9.2 × 105 m3 volume enters into the
lake with a maximum velocity of 34 m s− 1 (Fig. 7a and 7b). The
maximum flow velocity is 59 m s− 1 (212 km h− 1). The flow speed is
double to that of the flow speeds of the avalanche witnessed during
fieldwork. The modeled impulse wave generated at the site of impact
approaches the frontal moraine at a speed of 13 m s− 1 (Fig. 7c). The
wave overtops the frontal moraine with a flow height of 6.2 m draining a
total volume of 9.3 × 104 m3. The second avalanche (Avalanche-2), with
its source from another weak zone adjacent to the source of Avalanche-
1, carries 9.1 × 105 m3 of the total avalanche at a speed of 32.9 m s− 1 to
the lake. The resultant impulse wave overtops the frontal moraine with a
flow height of 5.7 m and an overtopping velocity of 12.4 m s− 1 (Fig. 7c)
draining 8.4 × 104 m3 of the lake volume. Although the model accounts
for the water depth at the impact site, the calculation does not include
the tsunami-type ramping up of wave amplitude and wave slow-down
due to lake shallowing toward the moraine, but overall the modeling
highlights the potential for large avalanches to generate powerful
waves. It is clear from the avalanche simulations that the frontal
moraine of Lower Barun Lake is susceptible to potential impact-waves
that can lead to the overtopping and subsequent breaching, following
which we evaluated the potential GLOFs of different magnitudes origi­
nating from the lake. We note that the higher magnitude events can also
originate from these slopes and that the overtopping heights of our
modeled waves are very small compared to some historic ones, such as
that of a GLOF in Peru (Hubbard et al., 2005; Emmer, 2017). The
overtopping hydrograph calculated for Avalanche-1 (impact-wave) has
a peak of 9298 m3s− 1 with a total overtopping duration of 21 s (Fig. 8).
For Avalanche-2 impact, the peak of 8300 m3s− 1 is calculated where the
overtopping duration is 20 s. The overtopping hydrographs were routed
downstream, revealing Df reaching up to ~ 8 m and Vf reaching up to 25
ms− 1. Both the flood waves (Avalanche-1 and Avalanche-2) travel 18 km
downstream of the lake, with overtopping from Avalanche-1being
slightly of higher magnitude. The overtopping flood does not affect
Fig. 8. Spatially distributed flow depths (Df) flow velocities (Vf) in an overtopping flood resulting from the impact of Avalanche-1 (a,b) and Avalanche-2 (c,d) on
Lower Barun Lake.
11
Journal of Hydrology 598 (2021) 126208
Barun Bazaar. At Yangle Kharke, Df and Vf reaches up to 1.6 m and 1.4
ms− 1, respectively. These modeled events are thus likely to be a rela­
tively high-frequency, low-magnitude type.
5.3. Moraine-breach and inter-comparison of GLOF discharge
GLOF hazard assessments of Lower Barun Lake were conducted
based on eight potential moraine-breach events that vary dimensionally
as a function of hb, Vw, Tf, and Bw. The physics and dynamics of the
breach’s erosional process are not modeled. Each of the potential breach
events produced a different outflow peak-discharge (Qmax) based on the
given breach parameters (Fig. 9). In general, the future GLOFs resulted
in higher discharges than the present-lake GLOFs due to increased lake
water volume. Here we assume that breach depths can reach a maximum
of 104 m in any (present and future) extreme-magnitude GLOF, if breach
incision completely propagates down to the base of the moraine (Fig. 4
and Fig. 6a). Such events are rare in nature and depend on the moraine
conditions and trigger magnitude. In the case of Lower Barun Lake, it
potentially drains the volume in the water column above 104 m, i.e., the
lake’s maximum drainable flood volume. For example, the potential
extreme-magnitude GLOF event (Scenario-1 present and scenario-5
future) produced a Qmax of 44,240 m3 s− 1 and 52,344 m3 s− 1, respec­
tively, immediately downstream of the lake (Fig. 9). A sharp rise is seen
in the outflow hydrograph where Qmax is reached within 20 min after the
initiation of the breach in both present and future cases, releasing a total
flood volume of 101.29 × 106 m3 and 179 × 106 m3, respectively
(Table 1). For other present scenarios (2–4), flood volume varies be­
tween 16 and 78.8 × 106 m3 while Qmax is reached 28–64 min after the
initiation of the breach in the present case. Similarly, for future scenarios
(6–8), the peak is reached 26–58 min after the breach event initiation as
it releases a flood volume of 26.9 to 139.5 × 106 m3.
5.4. Hydraulic characterization of GLOFs along the flow channel
GLOF inundations were calculated for eight different dam breach
scenarios with variations in their breach formation time (Tf), breach
height (hb), and breach width (Bw). We also evaluated Q (m3 s− 1), Vf (m
s− 1), and Df (m) by flood routing of the individual breach hydrographs
 A. Sattar et al.
12

Journal of Hydrology 598 (2021) 126208

Fig. 9. Schematic showing the breach parameters (Froehlich, 1995b) for different GLOF scenarios and their corresponding outflow hydrographs; Scenario-1 and Scenario-5 represents the extreme-magnitude GLOFs,
Scenario-2 and Scenario-6 represent the high-magnitude GLOFs, Scenario-3 and Scenario-7 represent the moderate-magnitude GLOFs, and Scenario- 4 and Scenario-8 represents the low-magnitude GLOFs of Lower
Barun Lake.
A. Sattar et al. Journal of Hydrology 598 (2021) 126208

obtained for each of the scenarios (Section 5.3; Fig. 9). This enabled us to

1530
characterize and compare the effects of flood-water progression from the

SC-8

925
440
372
NR
NR
various breach scenarios down the Barun and Arun Valley. Here we

5621
4266
3028
2331
1130
present the results of an inter-comparison of the flow hydraulics at six

SC-7

NR
different sites downstream of the lake (Barun Bazaar, and Site-1 to Site-

14552
12596
5; Fig. 1). Due to the large magnitude of the GLOF flows, any additional

9377
8149
4762
2587
SC-6
flow added by existing stream flows in the Barun and Arun rivers were
considered negligible.

26034
22063
16282
14613
8784
5472
SC-4 SC-5
5.4.1. Inter-comparison of routed hydraulics of potential current-GLOFs

261
100
NR
NR
NR
NR
In Scenario-1 (present extreme-magnitude scenario), the GLOF wave
first reaches Barun Bazaar 1 h 5 min after the initiation of the breach

Discharge (Q) (m3s− 1)

3198
2236
1464
SC-3

995
359
NR
event. For scenarios (2–4), the arrival time at this site ranges between 1
h 12 min to 6 h 30 min. Here, Df reaches up to 26 m with a maximum Vf

10988
8846
6298
5229
2943
1322
SC-2
of 9 m s− 1 (Table 2; Figs. 10, 11, and 12). The GLOFs could potentially
inundate a cluster of small houses located along the river-left bank of the

20810
17111
12522
10858
6413
3749
Arun River (Fig. 13). The low-magnitude GLOF (Scenario-4) terminates

SC-1 SC-2 SC-3 SC-4 SC-5 SC-6 SC-7 SC-8 SC-1 SC-2 SC-3 SC-4 SC-5 SC-6 SC-7 SC-8 SC-1 SC-2 SC-3 SC-4 SC-5 SC-6 SC-7 SC-8 SC-1
before it reaches Site-2 and moderate-magnitude GLOF (Scenario-3)
terminates before reaching Site-5.

NR
NR
<1
<1
4
3
At Site-1, the GLOFs potentially inundates a small settlement con­
sisting of 10–15 houses and a bridge with Df and Vf reaching up to 22 m

1.2
NR
<1
6
3

2
and 6 m s− 1. Farther downstream, Site-2 and Site-3 (located ~1 km

1.6
2.4
apart) reveal similar flow behavior, where Df ranges from 20 m to 27 m

Maximum flow velocity (Vf) (ms− 1)

9
5
1
3
and Vf between 1 m s− 1 to 3 m s− 1. Here, three small settlements, one

4.4
bridge, and 0.04 km2 of agricultural land become inundated (Fig. 13).

11
7
3
5
7
Among all the potential current GLOFs and the studied sites downstream

NR
NR
NR
NR
(Site-1 to Site-5), including Barun Bazaar, a maximum Df of 36 m

3
2
(Scenario-1) is reached in Site-4, where the 900 MW Arun III Hydro­

0.7
NR
<1
5
2

1
power project has its starting tunnel. This is due to the sharp turn of the

Flow hydraulics at different sites along the flow channel. The scenario that does not reach a specific site is given as “NR”.
river channel towards the west. Here, the potential GLOFs have a major

1.2
9
4
1
2

1
impact inundating a moderate-sized settlement, two bridges, and several
roads (Fig. 13). As the potential GLOFs propagate downstream, they

9
6
1
3
2
3
reach Site-5 (70 km downstream) with arrival timing of 4 h 48 min

NR
NR
(Scenario-1) and 7 h 8 min (Scenario-2), where Df reaches up to 19 m.

13

12
9

4
Here, the GLOF wave has a maximum Vf of 3 m s− 1, though its flow

NR
13
19
11
18
25
energy attenuates as it propagates farther downstream.
24
24
17
23
30
19
5.4.2. Routed hydraulics of potential future-GLOFs
The hydrodynamic routing and flow assessments of the future GLOFs
Maximum flow depth (Df) (m)

29
30
29
30
37
24
were evaluated for Scenarios 5, 6, 7, and 8 (Table 2). The routed
hydrographs were calculated based on the lake’s future extent and
NR
NR
NR
NR
2
3

volume (193.5 × 106 m3; see section 4.1). In a potential extreme-

magnitude future GLOF event (Scenario-5), the drainable volume (Vw)
NR
11
17
10
15
24

from the lake is calculated to be 179.0 × 106 m3 considering hb = 104 m

as a maximum breach depth (Table 1; Fig. 9). The flood wave arrives at
22
22
14
22
25
17

Barun Bazaar a few minutes (~5 min) earlier relative to that of Scenario-
26
27
20
27
36
19

1 (present extreme-magnitude GLOF). Here, the Df and Vf reach up to 29

m and 11 m s− 1, respectively (Table 2; Fig. 12). This accounts for an
3.0
3.6
5.5
6.0
NR
NR

increase of ~11 and ~20%, respectively, when compared to a present

extreme-magnitude event. The at-risk infrastructure remains the same at
2.07
2.33
3.07
3.23
4.93
NR

Site-1, 2, and 3; however, the flow hydraulics are variable (Figs. 10–12).
1.37
1.53
1.97
2.07
2.93
5.73

At Site-4, Df reaches a maximum of 37 m with a peak Vf of 7 m s− 1. This is

the highest velocity calculated at this site compared to all the other
1.01
1.20
1.53
1.60
2.27
4.33
Flood wave arrival (Time in h)

modeled GLOF events in this study. Here the inundation limits extend
further southward, engulfing the maximum area containing infrastruc­
6.5
9.2
NR
NR
NR
NR

ture and settlements. The flood wave travels further down the valley for
about ~3 km but has no impact due to the absence of infrastructure or
2.33
2.67
3.73
4.00
6.87
NR

settlement.
In Scenario-6, which represents a future high-magnitude GLOF, the
1.43
1.60
2.10
2.23
3.30
7.13

flood wave encounters Barun Bazaar 1 h 22 min after the breach event
Barun bazaar 1.07
1.20
1.60
1.67
2.43
4.83

(Table 2). Flow hydraulics indicate peak Df and Vf are decreased by 4 m

and 2 m s− 1, respectively, and its arrival time at Site-1 is delayed by 20
min compared to Scenario-5 (Fig. 12). At Site-2 and Site-3, the flow
hydraulics show a reduction of Df and Vf by an average of 9.5 m and 2 m
Table 2

Site-1
Site-2
Site-3
Site-4
Site-5

s− 1 at these sites compared to the extreme-magnitude future-lake GLOF.

The flood wave arrives at Site-4 40 min later compared to an extreme-

13
A. Sattar et al. Journal of Hydrology 598 (2021) 126208

Fig. 10. Scenario-based spatially distributed flow depths (Df) of Lower Barun GLOFs at different site locations (see Fig. 1 for site locations).

14
A. Sattar et al. Journal of Hydrology 598 (2021) 126208

Fig. 11. Scenario-based spatially distributed flow velocities (Vf) of Lower Barun GLOFs at different site locations (see Fig. 1 for site locations).

15
A. Sattar et al. Journal of Hydrology 598 (2021) 126208

Fig. 12. GLOF scenario-based discharge vs. time, flow depth vs. time, and flow velocity vs. time at each site downstream of Lower Barun Lake.

16
 A. Sattar et al.
17

Journal of Hydrology 598 (2021) 126208

Fig. 13. Google Earth images showing the potential inundation limits in the present and the future GLOFs; cross-sectional profiles across the flow channel at each selected site showing the present and future GLOF-water
surface elevation.
A. Sattar et al.
magnitude GLOF (Scenario-5) event and indicates a decrease in Df and Vf
by 7 m and 5.4 m s− 1, respectively. At Site-5, the flood wave is delayed
by 1 h 24 min when compared to Scenario-5. The overall attenuation of
Q from Barun Bazaar to Site-5 is 11,965 m3 s− 1.
In the case of a moderate-magnitude future-GLOF (Scenario-7), the
arrival of the flood wave to Barun Bazaar is slower by approximately an
hour compared to the extreme-magnitude Scenario-5. The flood wave
does not reach Site-5 as it loses its flow energy downstream. The at-risk
infrastructure upstream of Site-4 remains similar to the above scenarios.
At Site-4, Vf reduces to 1.2 m s− 1 with a discharge of 1130 m3 s− 1. Here,
however, Df rises to 25 m as the flow is obstructed due to the sharp bend
in the channel. The If decreases by 50% and 33% when compared to
Scenario-5 and Scenario-6, respectively.
The hydraulic evaluation of Scenario-8, a low-magnitude GLOF
event, reveals a similar GLOF wave to that of Scenario-4, though with a
comparable reduction in the peak flood by 592 m3 s− 1. It releases ~9%
of the total volume of the lake. The flood attenuates 42 km downstream
of the lake before it reaches Site-4. The total If along the channel is 6.6
km2 with mean Df and Vf values of 6 m and 5.6 m s− 1, respectively.
Of the eight GLOF scenarios described above (Section 5.3), only four
scenarios, including the potential extreme-magnitude GLOFs (Scenarios-
1, and -5), high-magnitude GLOFs (Scenarios-2, and -6) impact all the
studied sites (Barun Bazaar, Site-1 to 5) (Figs. 10–13). The moderate-
magnitude GLOFs has an impact up to Site-4, as it loses momentum
further downstream before reaching Site-5. The low magnitude present-
lake event impacts only until Site-1, whereas in the future-lake scenario,
impacts down to Site-3.
6. Discussion
6.1. Overtopping hazard
Overtopping flows occur when there is a sudden influx of mass into
the lake that displaces the water, forming displacement waves (Clague
and Evans, 2000; Westoby et al., 2014). At Lower Barun Lake, the po­
tential overtopping waves that resulted due to avalanche impacts are
very short-lived (Fig. 8); the hydrographs reached higher peaks draining
a small volume of the lake than that of the breach events (Table 1). Based
on the downstream flow hydraulics of the overtopping waves origi­
nating from the lake, it is seen that the hazard is minimal for down­
stream sites as the flood waves attenuate 18 km from the lake. The valley
is currently devoid of any settlement until Barun Bazaar, located 29.7
km downstream of the lake. However, at Yangle Kharka, flood waves
could potentially affect a couple of existing structures, though mini­
mally, since flow depth and flow velocity reach up to only 1.6 m and 1.4
ms− 1 (Fig. 8). The bathymetry of the lake (deepest in the west and
shoaling to the east) and emptying into a funnel-like outlet would pro­
mote the role of tsunami ramp-ups, whose energetic waves could be
focused down into the area of ponds and the outlet channel, potentially
incising the outlet and allowing a greater outflow. Our calculations show
that relatively small avalanches can lead to the ovetopping of the lake.
6.2. Plausibility of ‘slow GLOF’ and ‘fast GLOF’ scenarios
The field-collected bathymetric data allows realistic and robust
modeling of glacial lake outburst flood events in the Himalaya. We
modeled eight potential scenarios, though we recognize that GLOFs
could occur in several other ways. For example, the GLOF that occurred
from Peruvian Lake Palcacocha in 1941 was a ‘to the bottom’ breach
leading to almost 95% lake water released. However, this is uncommon
and only associated with specific topographical settings (Emmer, 2017).
Not all GLOFs are similar even if they occur from the same lake (Westoby
et al., 2014; Rounce et al., 2016). For example, in some circumstances, a
large fraction of a large glacial lake’s volume could drain out in a matter
of hours or days and still not cause extreme damage downstream except
to very low-lying infrastructure. We thus define in our discussion, “fast
18
Journal of Hydrology 598 (2021) 126208
GLOFs” whose proximal hydrographs peak within minutes of initiation
of an outburst event, and “slow GLOFs” whose hydrographs peaks hours
after an event starts and may take much longer to complete. A “slow
GLOF” may be generated when the drainable volume is constrained by
smaller breach dimensions or a large but slowly (hours-long) eroding
breach. On the contrary, “fast GLOFs” have much higher peak volume
discharge and potential flow energy per unit time as larger and fast-
formed breach dimensions allow for an enormous and rapid outlet-
discharge. Additionally, a GLOF might initiate slowly but then enter a
speedier phase as some critical failure occurs. Hence, “fast” and “slow”
GLOFs are not immutable. It is a concept that we introduce here to
propel focused discussion. So stated, “fast GLOFs” appear to pose a
relatively larger threat to downstream settlements and infrastructures,
as demonstrated by our modeling.
The rapidity of a moraine dam’s breakup depends on the magnitude
of the trigger and the initial disturbance that starts a self-accelerating
outflow. It also depends on the amount and distribution of buried ice,
the shape of the moraine, the hydraulic pressure at the base of the
moraine (hence, the depth of impounded water and height of the
moraine), and - an important factor that is rarely considered - the slope
and roughness of the bedrock beneath the moraine. The Lower Barun
moraine dam is high, and the lake is deep, while a shoaling near the
moraine could allow large tsunami runups. Ice and large boulders are
present in the moraine (Supplementary-1), but how much of these and
their internal arrangements are unknown, as is the shape and slope of
the bedrock base beneath the moraine. An assessment of the stability (or
metastability) of the moraine dam, or a numerical study of the dynamics
of a moraine erosional event, would require geophysical knowledge of
the moraine’s properties. Lacking knowledge of these, we have focused
on what is known.
The lake volume from the 2015 field survey was used to model po­
tential GLOFs in the current state. With the accelerating retreat of
connected Lower Barun Glacier, the lake area grew by 0.25 km2 in the
last 5 years (2015–2019) (Fig. S4). This implies that the GLOF hazard
will likely change as the melting and retreating glacier fills the enlarging
moraine-dammed proglacial lake (Clague and Evans, 1994). The seis­
mically active regime of the Nepal Himalaya has witnessed numerous
mass movement events like landslides, avalanches, and rockfalls of
various magnitudes (Kargel et al., 2016). These primary events (as
witnessed and identified in this study) may act as potential GLOF trig­
gers in the region, especially in the case of Lower Barun Lake, which is
surrounded by steep slopes with hanging ice masses. The Langmale
GLOF in 2017 (Byers et al., 2018) is an example of such an event that
occurred near Lower Barun Lake (Fig. 1). In fact, a landslide— maybe a
small moraine collapse— placed directly into the outlet channel,
potentially raising the lake by several meters over a several day period,
could prove to create a hazardous GLOF if that dam breaks up suddenly.
A low-magnitude avalanche event was witnessed and captured dur­
ing our fieldwork at the lake site in 2015 (Fig. S5), but it did not result in
any known unusual discharge event. However, our modeling of rela­
tively higher magnitude avalanches shows that the frontal moraine of
Lower Barun Lake is susceptible to overtopping. The GLOF hazard
magnitudes due to avalanches and rockfalls may extend across the full
range of the potential GLOF scenarios modeled in this study, the smaller
ones being more frequent than the larger. The peak discharge from
Lower Barun GLOFs at the current state can reach as high as 20,810 m3
s− 1 and can increase by 5224 m3 s− 1 in the future. Flow depths and
velocities reaching up to 29 m and 11 m s− 1 can present the valley with
great risk. Possibly the typical risk would be to downstream populations
along the channel.
The breach depths of the moraine in the two low-magnitude events
categorized as slow GLOFs account for only 15% of the total moraine
height of the lake, thereby restricting the lake release volume and
discharge rate. Such flow events pose no severe threat to the Barun
Valley as the flow energy is significantly reduced as it propagates
downstream and largely attenuates into a steady flow. These events may
A. Sattar et al.
be common in the Himalaya, although likely remain unreported as these
floods typically have little or no significant impact on human settle­
ments or infrastructure (Veh et al., 2018).
In a dam breach, the opening forms at the top of the embankment
barrier, and as the failure continues, it gradually grows. A breach may
also develop as a piping failure, including openings or holes that prop­
agate internal erosion within an embankment and grow outwardly until
the opening reaches the top of the embankment (Westoby et al., 2015).
Regardless if a breach forms due to embankment overtopping or piping,
the size of the opening will increase over time as more and more water
passes over or through the opening. Once initiated, a GLOF may progress
to an extreme-magnitude event (Huggel et al., 2004), such as scenarios 1
and 4 in this study, if the overtopping wave or the ensuing outflow
continues to erode the dam progressively. Therefore, the hazard
assessment in the case of a potential extreme-magnitude GLOF cannot be
neglected.
By contrast, our evaluation of flood wave behavior for extreme-
magnitude GLOFs, high-magnitude GLOFs, and moderate-magnitude
GLOFs (both present and the future scenarios) ties them to the produc­
tion of fast-GLOFs; these hazard events have much higher flow energy
compared to the slow-GLOFs. This is because the GLOF peak is reached
more rapidly during a progressive breaching event with breach depths
sometimes propagating down to the base of the frontal moraine (104 m
in case of Lower Barun). Moreover, the breach widths (Bw) are also
significantly higher (3 to 4 times) than that of the slow GLOFs allowing
higher outflow discharge from the lake. The flow velocities (Vf) along
the routed channel are comparatively much higher and therefore travel
downstream for longer distances, both values an order of magnitude
greater when compared to the slow events.
GLOF triggers and breach formation mechanisms remain uncertain
in most cases; however, detailed engineering field investigations can
reduce them. This is most often very challenging in harsh terrain such as
the Himalaya. However, for disaster preparedness and mitigation, hy­
draulic modeling of these events provides valuable initial insights and
motivations for the needed geophysical field studies and engineering
work.
In the aforementioned case of Peru’s Palcacocha GLOF, the transport
and deposition of giant boulders and finer sediment caused thousands of
deaths; casualties would likely have been far smaller if the event only
involved water. The amount of ingested, transported, and deposited
sediment from possible GLOFs from Lower Barun Lake will be tightly
connected to the duration of the peak discharge part of the flood emis­
sion, the eroded volume of the moraine dam, and the amount of erodible
sediments along the channel. Hence, the sediment component is also
tied to whether it is a fast or a slow GLOF.
6.3. Shifting Lower Barun hazard in the future
The Barun-Arun valley, geographically located at the base of Mt.
Makalu, has been a favorable region for hydropower generation and
currently hosts several infrastructure and hydropower plants that are
either fully functional, under construction, or planned. These new de­
velopments are opening ways for the communities to expand in the
valley. In addition to the enormous hydropower potential the valley
possesses, it also hosts several glacial lakes in the high altitude glaci­
erized regions, of which Upper Barun and Lower Barun lakes are the
prominent ones (Fig. 1), but as the Langmale flood shows, are not the
only dangerous lakes. These high altitude lakes and potential GLOFs
present enough downstream threat to impact or destroy the at-risk
infrastructure; these potential impacts are intensifying both due to in­
creases in the frequency of hazards and vulnerabilities with increased
development. Currently, the valley is not equipped with any mitigation
measures to counter these threats, and development planning proceeds
as though the risk is nil. This study provides the basis for GLOF pre­
paredness based on the present and future anticipated GLOFs in the
valley. The modeled GLOFs were compared for flow hydraulics along the
19
Journal of Hydrology 598 (2021) 126208
flow channel to evaluate the shift in hazard potential from the current
state to future state. Of the eight potential GLOF events we have
modeled, the slow (low-magnitude) GLOF events (Scenario-4 and
Scenario-8) show similar behavior in both present and future states due
to their limited breach and drainable flood volumes. However, the
extreme-magnitude cases (Scenario-1 and Scenario-5) reveals a change
in flow dynamics and GLOF hazards along the Lower Barun valley. An
overall rise in the total inundation modeled in a potential future-lake
GLOF event (Scenario-5, 6, and 7) is observed when compared to the
current scenario (Scenario-1, 2, and 3) (Figs. 10, 11 and 13). The flow
hydraulics Q, Vf, and Df comparison for the present and future scenarios
shows that the Q and Vf shift is more profound than Df (Fig. 12). This can
be attributed to the higher flow energy and discharge, which tends to
reduce the frictional resistance compared to a relatively lower discharge
event. An increase in Q and Vf by 33% and 36%, respectively, are
observed in the future extreme-magnitude event compared to a present
extreme-magnitude GLOF. Further, the arrival timings of the GLOF wave
at the selected sites decrease in a potential future event due to higher
PFV, which results in a higher momentum. The attenuation of the flood
wave is seen to be prominent more towards the downstream part of the
valley after the flood crosses Site-3, where the difference in flood wave
arrival timing increases (Fig. S7). This is due to higher channel friction
and lower flow volume as the GLOF propagates downstream. This in­
formation becomes vital in designing and implementing any adaptation
measures along the Barun and Arun rivers’ flow path as part of disaster
preparedness.
Furthermore, a comparison of the flow hydraulics revealed that the
GLOF arrival at the first site of interest (Barun Bazaar) decreases by
4–16 min in the fast GLOFs. This is due to the comparatively more rapid
peaking of Vf in future GLOFs, which is governed by the breach pa­
rameters and PFV. This difference in GLOF arrival timing is observed
more prominently in the downstream part of the valley (after Site-3).
This may be attributed to the loss of flow volume and channel resis­
tance as the flood propagates down the valley. The decreased time
carries with it a greater challenge for potential GLOF warning systems
and evacuation scenarios. In addition, the total If along the flow channel
increases by ~15% in case of future GLOFs, compared to potential
present events. This is due to the rise of Df, resulting in higher water-
surface elevation in potential future GLOFs (Fig. 13). Although the
GLOF hazard in both present and future scenarios are very high, a future
event coupled with increasing development intensifies the risk that
Lower Barun Lake will pose to the low-lying areas downvalley.
6.4. Future directions
We do not model the erosion dynamics of the breach in the moraine
dam. Yager et al. (2012) modeled grain transport in steeply sloping
mountain channels and emphasized the important hydrodynamic role of
large, immobile protruding obstacles—like large boulders— in reducing
the transport of smaller mobile grains. In the case of Lower Barun Lake’s
end moraine dam, there is a ubiquitous presence within a fine-grained
matrix of large boulders— many are 1 to 6 m across (Supplementary
section-1), and some are several tens of meters. Whereas 1-m boulders
are normal bedload material and can be readily transported in mountain
streams along with finer particles over time, they could be transported
rapidly during high discharge GLOFs. The larger boulders could remain
immobile and retard erosional downcutting such that peak GLOF
discharge will be limited, or alternatively an incredible outflow may
rapidly mobilize the large boulders and prevent them from armoring the
channel. A question for Lower Barun Lake is whether there are enough
large boulders in the dam to create an armoring lag deposit. In principle,
but not modeled here, a large initial outflow could move large boulders
and induce runaway erosion and outflow. Future work should link the
overtopping and erosion dynamics, including the critical roles of large
boulders and the erosional susceptibility of heterogeneously composed
moraine dams, to understand the full process chain, which we have not
A. Sattar et al.
done currently in this paper. To undertake this work we would need
detailed in situ geophysical and sedimentological data on the moraine,
more detailed bathymetry of the lakes, and then theoretical modeling
(analytical or numerical) to understand the processes and threshold
flows under which the moraine could undergo runaway erosion. Also
needed is a better knowledge of the geotechnical properties of poten­
tially unstable masses of rock and ice on the mountain slopes, their
potential failure mechanisms and mass flow behavior. Overall, an
obvious future direction is to include sediment explicitly in the hy­
draulic modeling because the sediment will affect the flow depth, arrival
time at various points, and deposition at downstream locations. It may
thus become useful, in the next step of the investigation, to distinguish
and model “dirty GLOFs” (debris-rich) and “clean GLOFs” (mainly
water) with an emphasis on viscous flows.
7. Conclusions
Detailed hydrodynamic scenario modeling reveals the hazard po­
tential of moraine-dammed Lower Barun Lake, the deepest and one of
the largest known glacial lakes in Nepal. The lake has shown substantial
area growth from 0.04 km2 in 1975 to 2.09 km2 in 2019. The ‘present’
(2015) volume of the lake was measured as 112.3 × 106 m3 and the
future volume is calculated as 193.5 × 106 m3. The flow hydraulics of
the overtopping waves caused due to impact from potential avalanches
show that the Barun valley is minimally impacted as the upstream part
of the valley is devoid of infrastructure. In the case of GLOFs, the
maximum drainable volumes in an extreme-magnitude flood originating
from the lake in the present and future are calculated to be 101.2 × 106
m3 and 179 × 106 m3, respectively. Additionally, we evaluated hydro­
dynamic flow metrics for sites along the river valley for potential GLOFs
originating from both present and the estimated future lake volume. The
at-risk infrastructure remains similar in the case of present and future
“fast GLOFs.” However, the impact changes in the future as flow hy­
draulics (Vf, If, and Df) increases in potential future events. This is due to
larger lake volume resulting in rapid breaching and higher outflow peak
and total discharge. Numerous small settlements and infrastructure,
including agricultural fields and hydroelectric power projects (under
construction or planned), are under potential flood risk. The possible
impact of “slow GLOFs” in the valley is significantly less than that of fast
events. As both the Lower Barun Lake and the Barun valley evolve, the
potential threat from this upstream lake increases. The magnitude of
‘fast GLOF’ events can significantly affect the downstream infrastructure
both in the current state and in the future, affecting up to 70 km
downstream of the lake. Based on the present study, regular monitoring
of Lower Barun Lake is highly recommended. Further, the results pre­
sented here can serve as base information for disaster and flood miti­
gation. Particularly needed are in situ geophysical studies that may
provide constraints on the potential for “fast” versus “slow” GLOFs.
CRediT authorship contribution statement
Ashim Sattar: Conceptualization, Methodology, Writing - original
draft, Writing - review & editing. Umesh K. Haritashya: Conceptuali­
zation, Methodology, Data curation, Writing - original draft, Writing -
review & editing. Jeffrey S. Kargel: Data curation, Writing - review &
editing. Gregory J. Leonard: Data curation, Writing - review & editing.
Dan H. Shugar: Writing - review & editing. Donald V. Chase: Meth­
odology, Writing - review & editing.
Declaration of Competing Interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
20
Journal of Hydrology 598 (2021) 126208
Acknowledgments
We thank the associate editor and two anonymous reviewers for their
insightful comments. The authors would like to acknowledge USGS,
DigitalGlobe, Planet, Google Earth, Alaska Satellite Facility for various
remote sensing data used in the study. We acknowledge the US Army
Corps of Engineers for HEC-RAS model. We are grateful for the support
from various National Aeronautics and Space Administration (NASA)
funding programs for the last several years. Specifically, the following
NASA grants made this work possible at different stages: HEC-RAS
modeling work under High Mountain Asia grant 80NSSC19K0653 (to
UKH, JSK, DHS) and Cryosphere grant 80NSSC20K1442 (to UKH and
JSK), RAMMS avalanche modeling under Interdisciplinary Research in
Earth Science grant 80NSSC18K0432 (to UKH and JSK), Lower Barun
field bathymetric data collection in 2015 under NASA-USAID SERVIR
grant NNX12AO96G (to UKH, JSK, GJL).
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.jhydrol.2021.126208.
References
Allen, S.K., Rastner, P., Arora, M., Huggel, C., Stoffel, M., 2016. Lake outburst and debris
flow disaster at Kedarnath, June 2013: hydrometeorological triggering and
topographic predisposition. Landslides 13 (6), 1479–1491.
Anacona, P.I., Mackintosh, A., Norton, K.P., 2015a. Hazardous processes and events from
glacier and permafrost areas: lessons from the Chilean and Argentinean Andes. Earth
Surf. Proc. Land. 40 (1), 2–21.
Anacona, P.I., Mackintosh, A., Norton, K., 2015b. Reconstruction of a glacial lake
outburst flood (GLOF) in the Engaño Valley, Chilean Patagonia: Lessons for GLOF
risk management. Sci. Total Environ. 527, 1–11.
Bajracharya, S.R., Mool, P., 2009. Glaciers, glacial lakes and glacial lake outburst floods
in the Mount Everest region. Nepal. Ann. Glaciol. 50 (53), 81–86.
Bajracharya, S.R., Shrestha, B.R., 2011. The status of glaciers in the Hindu Kush-
Himalayan region. International Centre for Integrated Mountain Development
(ICIMOD).
Bartelt, P., Buehler, Y., Christen, M., Deubelbeiss, Y., Salz, M., Schneider, M., and
Schumacher, L.: RAMMS: Rapid Mass Movement Simulation: A numerical model for
snow avalanches in research and practice. User Manual v1.5 – Avalanche, Swiss
Federal Institute for Forest, Snow and Landscape Research WSL, Birmensdorf, 2013.
Bolch, T., Kulkarni, A., Kääb, A., Huggel, C., Paul, F., Cogley, J.G., Frey, H., Kargel, J.S.,
Fujita, K., Scheel, M., Bajracharya, S., 2012. The state and fate of Himalayan glaciers.
Science 336 (6079), 310–314.
Byers, A.C., Rounce, D.R., Shugar, D.H., Lala, J.M., Byers, E.A., Regmi, D., 2018.
A rockfall-induced glacial lake outburst flood, Upper Barun Valley, Nepal. Landslides
16 (3), 533–549.
Byers, A.C., Chand, M.B., Lala, J., Shrestha, M., Byers, E.A., Watanabe, T., 2020.
Reconstructing the history of glacial lake outburst floods (GLOF) in the
Kanchenjunga Conservation Area, East Nepal: an interdisciplinary approach.
Sustainability 12 (13), 5407.
Campbell, J.G., Pradesh, H., 2005. Inventory of Glaciers, Glacial Lakes and the
Identification of Potential Glacial Lake Outburst Floods (GLOFs) Affected by Global
Warming in the Mountains of India, Pakistan and China/Tibet Autonomous Region.
International Centre for Integrated Mountain Development, GP O Box, p. 3226.
Carey, M., Huggel, C., Bury, J., Portocarrero, C., Haeberli, W., 2012. An integrated socio-
environmental framework for glacier hazard management and climate change
adaptation: lessons from Lake 513, Cordillera Blanca. Peru. Climatic Change 112
(3–4), 733–767.
Christen, M., Kowalski, J., Bartelt, P., 2010. RAMMS: Numerical simulation of dense
snow avalanches in three-dimensional terrain. Cold Reg. Sci. Technol. 63 (1–2),
1–14.
Clague, J.J., Evans, S.G., 1994. Formation and Failure of Natural Dams in the Canadian
Cordillera, 464. Geological Survey of Canada.
Clague, J.J., Evans, S.G., 2000. A review of catastrophic drainage of moraine-dammed
lakes in British Columbia. Quaternary Sci. Rev. 19 (17–18), 1763–1783.
Clague, J.J., O’Connor, J.E., 2015. Glacier-related outburst floods. In Snow and ice-
related hazards, risks and disasters. Academic Press, pp. 487–519.
Cook, K.L., Andermann, C., Gimbert, F., Adhikari, B.R., Hovius, N., 2018. Glacial lake
outburst floods as drivers of fluvial erosion in the Himalaya. Science 362 (6410),
53–57.
Cuellar, A.D., McKinney, D.C., 2017. Decision-making methodology for risk management
applied to Imja lake in Nepal. Water 9 (8), 591.
Dahal, P.R., Paudyal, K.R., Rajaure, S., 2018. Geophysical study on moraine dam of Imja
Glacial Lake in Eastern Nepal using Electrical Resistivity Tomography Method.
J. Nepal Geol. Soc. 55 (1), 15–22.
A. Sattar et al.
Dhote, P.R., Aggarwal, S.P., Thakur, P.K., Garg, V., 2019. Flood inundation prediction for
extreme flood events: a case study of Tirthan River, North West Himalaya.
Himalayan Geol. 40 (2), 128–140.
Emmer, A., Klimeš, J., Mergilid, M., Vilíme, V., Cochachin, A., 2016. 882 lakes of the
Cordillera Blanca: An inventory, classification, evolution and assessment of
susceptibility to outburst floods. CATENA 147, 269–279.
Emmer, A., 2017. Glacier retreat and glacial lake outburst floods (GLOFs). In Oxford
Research Encyclopedia of Natural Hazard Science.
Emmer, A., 2018. GLOFs in the WOS: Bibliometrics, geographies and global trends of
research on glacial lake outburst floods (Web of Science, 1979–2016). Nat. Hazards
Earth Syst. Sci. 18 (3), 813.
Evers, F., Heller, V., Fuchs, H., Hager, W.H., Boes, R., 2019. Landslide-generated Impulse
Waves in Reservoirs: Basics and Computation. VAW-Mitteilungen, 254.
Farinotti, D., Huss, M., Fürst, J.J., Landmann, J., Machguth, H., Maussion, F., Pandit, A.,
2019. A consensus estimate for the ice thickness distribution of all glaciers on Earth.
Nat. Geosci. 12 (3), 168–173.
Frey, H., Haeberli, W., Linsbauer, A., Huggel, C., Paul, F., 2010. A multi-level strategy for
anticipating future glacier lake formation and associated hazard potentials. Nat.
Hazards Earth Syst. Sci. 10 (2), 339–352.
Frey, H., Huggel, C., Chisolm, R.E., Baer, P., McArdell, B., Cochachin, A.,
Portocarrero, C., 2018. Multi-source glacial lake outburst flood hazard assessment
and mapping for Huaraz, Cordillera Blanca, Peru. Front. Earth Sci. 6, 210.
Froehlich, D.C., 1995a. Peak outflow from breached embankment dam. J. Water Resour.
Plann. Manage. 121 (1), 90–97.
Froehlich, D.C., 1995b. Embankment Dam Breach Parameters Revisited, Water
Resources Engineering, Proceedings of the 1995 ASCE Conference on Water
Resources Engineering, San Antonio, Texas.
Fujita, K., Sakai, A., Nuimura, T., Yamaguchi, S., Sharma, R.R., 2009. Recent changes in
Imja Glacial Lake and its damming moraine in the Nepal Himalaya revealed by in
situ surveys and multi-temporal ASTER imagery. Environ. Res. Lett. 4 (4), 045205.
Fujita, K., Sakai, A., Takenaka, S., Nuimura, T., Surazakov, A.B., Sawagaki, T.,
Yamanokuchi, T., 2013. Potential flood volume of Himalayan glacial lakes. Nat.
Hazards Earth Syst. Sci. 13 (7).
Haritashya, U.K., Kargel, J.S., Shugar, D.H., Leonard, G.J., Strattman, K., Watson, C.S.,
Shean, D., Harrison, S., Mandli, K.T., Regmi, D., 2018. Evolution and controls of
large glacial lakes in the Nepal Himalaya. Remote Sensing 10 (5), 798.
Harrison, S., Kargel, J.S., Huggel, C., Reynolds, J., Shugar, D.H., Betts, R.A., Emmer, A.,
Glasser, N., Haritashya, U.K., Klimeš, J., Reinhardt, L., 2018. Climate change and the
global pattern of moraine-dammed glacial lake outburst floods. Cryosphere 12 (4),
1195–1209.
Haeberli, W., Buetler, M., Huggel, C., Friedli, T.L., Schaub, Y., Schleiss, A.J., 2016. New
lakes in deglaciating high-mountain regions–opportunities and risks. Clim. Change
139 (2), 201–214.
Haeberli, W., Linsbauer, A., 2013. Brief communication“ Global glacier volumes and sea
level-small but systematic effects of ice below the surface of the ocean and of new
local lakes on land”. The Cryosphere 7 (3), 817–821.
Hanisch, J., Delisle, G., Pokhrel, A.P., Dixit, A.M., Reynolds, J.M., Grabs, W.E., 1998. The
Thulagi glacier lake, Manaslu Himal, Nepal*Hazard assessment of a potential
outburst. Proceedings EighthInternational Congress, International Association for
Engineering Geology and the Environment, 21}25 September, Vancouver, Canada,
pp. 2209–2215.
Hanisch, J., Pokhrel, A.P., Dixit, A.M., Grabs, W.E., Reynolds, J.M., 1999. GLOF
mitigation strategies*lessons learnt from studyingthe Thulagi glacier lake. Nepal
[abstract]. J. Nepal Geol. Soc. 20, 163.
Heller, V., Hager, W.H., 2010. Impulse product parameter in landslide generated impulse
waves. J. Waterw. Port Coastal Ocean Eng. 136 (3), 145–155.
Hubbard, B., Heald, A., Reynolds, J.M., Quincey, D., Richardson, S.D., Luyo, M.Z.,
Portilla, N.S., Hambrey, M.J., 2005. Impact of a rock avalanche on a moraine-
dammed proglacial lake: Laguna Safuna Alta, Cordillera Blanca, Peru. Earth Surface
Processes Landf. 30 (10), 1251–1264.
Huggel, C., Haeberli, W., Kääb, A., Bieri, D., Richardson, S., 2004. An assessment
procedure for glacial hazards in the Swiss Alps. Can. Geotech. J. 41 (6), 1068–1083.
Kargel, J., Furfaro, R., Kaser, G., Leonard, G., Fink, W., Huggel, C., Kääb, A., Raup, B.,
Reynolds, J., Wolfe, D., Zapata, M., 2011. In: ASTER Imaging and Analysis of Glacier
Hazards, Chapter 15 in Land Remote Sensing and Global Environmental Change:
NASA’s Earth Observing System and the Science of Terra and Aqua. Springer, New
York, pp. 325–373.
Kargel, J.S., Leonard, G.J., Shugar, D.H., Haritashya, U.K., Bevington, A., Fielding, E.J.,
Fujita, K., Geertsema, M., Miles, E.S., Steiner, J., Anderson, E., 2016. Geomorphic
and geologic controls of geohazards induced by Nepal’s 2015 Gorkha earthquake.
Science 351 (6269).
Kattelmann, R., 2003. Glacial lake outburst floods in the Nepal Himalaya: a manageable
hazard? Nat. Hazards 28 (1), 145–154.
Khanal, N.R., Mool, P.K., Shrestha, A.B., Rasul, G., Ghimire, P.K., Shrestha, R.B., Joshi, S.
P., 2015. A comprehensive approach and methods for glacial lake outburst flood risk
assessment, with examples from Nepal and the transboundary area. Int. J. Water
Resour. Dev. 31 (2), 219–237.
Khadka, N., Zhang, G., Thakuri, S., 2018. Glacial lakes in the Nepal Himalaya: inventory
and decadal dynamics (1977–2017). Remote Sensing 10 (12), 1913.
Kirschbaum, D., Watson, C.S., Rounce, D.R., Shugar, D., Kargel, J.S., Haritashya, U.K.,
Amatya, P., Shean, D., Anderson, E.R., Jo, M., 2019. The state of remote sensing
capabilities of cascading hazards over High Mountain Asia. Front. Earth Sci. 7, 197.
Klimeš, J., Benešová, M., Vilímek, V., Bouška, P., Rapre, A.C., 2014. The reconstruction
of a glacial lake outburst flood using HEC-RAS and its significance for future hazard
assessments: an example from Lake 513 in the Cordillera Blanca, Peru. Nat. Hazards
71 (3), 1617–1638.
21
Journal of Hydrology 598 (2021) 126208
Kropáček, J., Neckel, N., Tyrna, B., Holzer, N., Hovden, A., Gourmelen, N., Schneider, C.,
Buchroithner, M., Hochschild, V., 2015. Repeated glacial lake outburst flood
threatening the oldest Buddhist monastery in north-western Nepal. Nat. Hazards
Earth Syst. Sci. 15 (10), 2425–2437.
Kraaijenbrink, P.D.A., Bierkens, M.F.P., Lutz, A.F., Immerzeel, W.W., 2017. Impact of a
global temperature rise of 1.5 degrees Celsius on Asia’s glaciers. Nature 549 (7671),
257–260.
Lala, J.M., Rounce, D.R., McKinney, D.C., 2018. Modeling the glacial lake outburst flood
process chain in the Nepal Himalaya: reassessing Imja Tsho’s hazard. Hydrol. Earth
Syst. Sci. 22 (7), 3721–3737.
Lavell, A., Oppenheimer, M., Diop, C., Hess, J., Lempert, R., Li, J., Muir-Wood, R.,
Myeong, S., Moser, S., Takeuchi, K., Cardona, O.D., 2012. Climate change: new
dimensions in disaster risk, exposure, vulnerability, and resilience. In: Managing the
risks of extreme events and disasters to advance climate change adaptation: Special
report of the Intergovernmental Panel on Climate Change. Cambridge University
Press, pp. 25–64.
Linsbauer, A., Frey, H., Haeberli, W., Machguth, H., Azam, M.F., Allen, S., 2016.
Modelling glacier-bed overdeepenings and possible future lakes for the glaciers in
the Himalaya—Karakoram region. Ann. Glaciol. 57 (71), 119–130.
Linsbauer, A., Paul, F., Haeberli, W., 2012. Modeling glacier thickness distribution and
bed topography over entire mountain ranges with GlabTop: application of a fast and
robust approach. J. Geophys. Res. Earth Surf. 117 (F3).
Majeed, U., Rashid, I., Sattar, A., Allen, S., Stoffel, M., Nüsser, M. and Schmidt, S., 2020.
Recession of Gya Glacier and the 2014 glacial lake outburst flood in the Trans-
Himalayan region of Ladakh, India. The Science of the Total Environment,
pp.144008–144008.
Maskey, S., Kayastha, R.B. and Kayastha, R., 2020. Glacial Lakes Outburst Floods
(GLOFs) Modelling of Thulagi and Lower Barun Glacial Lakes of Nepalese Himalaya.
Progress in Disaster Science, p. 100106.
Meon, G., Schwarz, W., 1993. Estimation of glacier lake outburst flood and its impact on
a hydro project in Nepal. Int. Assoc. Hydrol. Sci. 218, 331–340.
Nie, Y., Sheng, Y., Liu, Q., Liu, L., Liu, S., Zhang, Y., Song, C., 2017. A regional-scale
assessment of Himalayan glacial lake changes using satellite observations from 1990
to 2015. Remote Sens. Environ. 189, 1–13.
Osti, R., Egashira, S., 2009. Hydrodynamic characteristics of the Tam Pokhari Glacial
Lake outburst flood in the Mt. Everest region, Nepal. Hydrol. Proc.: Int. J. 23 (20),
2943–2955.
Pant, S.R., Reynolds, J.M., 1999. Subsurface electrical imaging tech-niques for the
investigation of Thulagi Glacier Lake dam [abstract]. J. Nepal Geol. Soc. 20, 80.
Paul, F., Bolch, T., 2019. Glacier Changes Since the Little Ice Age. In: Geomorphology of
Proglacial Systems. Springer, Cham, pp. 23–42.
Regmi, D., Kargel, J.S., Leonard, G.J., Haritashya, U.K., Karki, A., Poudyal, S., 2017.
Potential GLOF Hazards and Initiatives taken to minimize its Impacts on
Downstream Communities and Infrastructures in Nepal Himalaya. AGUFM 2017,
C33A–1166.
Quincey, D.J., Richardson, S.D., Luckman, A., Lucas, R.M., Reynolds, J.M., Hambrey, M.
J., Glasser, N.F., 2007. Early recognition of glacial lake hazards in the Himalaya
using remote sensing datasets. Global Planet. Change 56 (1–2), 137–152.
Reynolds, J.M., Taylor, P.J., 2004. Inventory of Glaciers, Glacial Lakes and Glacial Lake
Outburst Floods, Monitoring and Early Warning Systems in the Hindu Kush-
Himalaya Region: Nepal.
Richardson, S.D., Reynolds, J.M., 2000. An overview of glacial hazards in the Himalayas.
Quat. Int. 65, 31–47.
Roe, G.H., Baker, M.B., Herla, F., 2017. Centennial glacier retreat as categorical evidence
of regional climate change. Nat. Geosci. 10 (2), 95–99.
Rounce, D.R., McKinney, D.C., Lala, J.M., Byers, A.C., Watson, C.S., 2016. A new remote
hazard and risk assessment framework for glacial lakes in the Nepal Himalaya.
Hydrol. Earth Syst. Sci. 20, 3455–3475.
Rounce, D.R., Watson, C.S., McKinney, D.C., 2017. Identification of hazard and risk for
glacial lakes in the Nepal Himalaya using satellite imagery from 2000–2015. Remote
Sensing 9 (7), 654.
Sattar, A., Goswami, A., Kulkarni, A.V., 2019a. Hydrodynamic moraine-breach modeling
and outburst flood routing – a hazard assessment of the South Lhonak lake, Sikkim.
Sci. Total Environ. 668, 362–378.
Sattar, A., Goswami, A., Kulkarni, A.V., Das, P., 2019b. Glacier-surface velocity derived
ice volume and retreat assessment in the Dhauliganga basin, Central Himalaya – a
remote sensing and modeling based approach. Front. Earth Sci. 7, 105.
Sattar, A., Goswami, A., Kulkarni, A.V., 2019c. Application of 1D and 2D hydrodynamic
modeling to study glacial lake outburst flood (GLOF) and its impact on a hydropower
station in Central Himalaya. Nat. Hazards 97 (2), 535–553.
Sattar, A., Goswami, A., Kulkarni, A., Emmer, A., 2020. Lake evolution, hydrodynamic
outburst flood modeling and sensitivity analysis in the Central Himalaya: a case
study. Water 12 (1), 237.
Shugar, D.H., Burr, A., Haritashya, U.K., Kargel, J.S., Watson, C.S., Kennedy, M.C.,
Bevington, A.R., Betts, R.A., Harrison, S., Strattman, K., 2020. Rapid worldwide
growth of glacial lakes since 1990. Nat. Clim. Change 1–7.
Shrestha, A.B., Eriksson, M., Mool, P., Ghimire, P., Mishra, B., Khanal, N.R., 2010. Glacial
lake outburst flood risk assessment of Sun Koshi basin, Nepal. Geomatics, Nat. Haz.
Risk 1 (2), 157–169.
Shukla, A., Garg, P.K., Srivastava, S., 2018. Evolution of glacial and high-altitude lakes in
the Sikkim, Eastern Himalaya over the past four decades (1975–2017). Front.
Environ. Sci. 6, 81.
Schneider, D., Huggel, C., Cochachin, A., Guillén, S., García, J., 2014. Mapping hazards
from glacier lake outburst floods based on modelling of process cascades at Lake
513, Carhuaz, Peru. Adv. Geosci. 35, 145–155.
Singh, V.P., 1996. Dam Breach Modelling Technology vol. 17, 242.
A. Sattar et al.
Somos-Valenzuela, M.A., McKinney, D.C., Byers, A.C., Rounce, D.R., Portocarrero, C.,
Lamsal, D., 2015. Assessing downstream flood impacts due to a potential GLOF from
Imja Tsho in Nepal. Hydrol. Earth Syst. Sci. 19 (3).
Somos-Valenzuela, M.A., Chisolm, R.E., Rivas, D.S., Portocarrero, C., McKinney, D.C.,
2016. Modeling a glacial lake outburst flood process chain: the case of Lake
Palcacocha and Huaraz, Peru. Hydrol. Earth Syst. Sci. 20 (6), 2519.
Thakur, P.K., Aggarwal, S., Aggarwal, S.P., Jain, S.K., 2016. One-dimensional
hydrodynamic modeling of GLOF and impact on hydropower projects in
Dhauliganga River using remote sensing and GIS applications. Nat. Hazards 83 (2),
1057–1075.
Uddin, K., Shrestha, H.L., Murthy, M.S.R., Bajracharya, B., Shrestha, B., Gilani, H.,
Pradhan, S., Dangol, B., 2015. Development of 2010 national land cover database for
the Nepal. J. Environ. Manage. 148, 82–90.
USAID, 2014. Affirmative investigations for hydropower projects in Nepal: upper
Marsyandi 2, upper Trisuli 1, and Upper Arun. U.S. Agency for International
Development’s Bureau forEconomic Growth. Education and Environment (USAID/
E3), Washington DC.
UNDP – United Nations Development Programme: Community Based Glacier Lake
Outburst and Flood Risk Reduction in Nepal, Project Document, UNDP
Environmental Finance Services, Kathmandu, Nepal, 2013.
Veh, G., Korup, O., Roessner, S., Walz, A., 2018. Detecting Himalayan glacial lake
outburst floods from Landsat time series. Remote Sens. Environ. 207, 84–97.
Veh, G., Korup, O., Walz, A., 2020. Hazard from Himalayan glacier lake outburst floods.
Proc. Natl. Acad. Sci. 117 (2), 907–912.
Vilímek, V., Klimeš, J., Emmer, A., Benešová, M., 2015. Geomorphologic impacts of the
glacial lake outburst flood from Lake No. 513 (Peru). Environ. Earth Sci. 73 (9),
5233–5244.
22
Journal of Hydrology 598 (2021) 126208
Vuichard, D., Zimmermann, M., 1987. The 1985 catastrophic drainage of a moraine-
dammed lake, Khumbu Himal, Nepal: cause and consequences. Mt. Res. Dev.
91–110.
Wahl, T.L., 2004. Uncertainty of predictions of embankment dam breach parameters.
J. Hydraul. Eng. 130 (5), 389–397.
Wang, W., Gao, Y., Anacona, P.I., Lei, Y., Xiang, Y., Zhang, G., Li, S., Lu, A., 2018.
Integrated hazard assessment of Cirenmaco glacial lake in Zhangzangbo valley,
Central Himalayas. Geomorphology 306, 292–305.
Watanbe, T., Rothacher, D., 1996. The 1994 Lugge Tsho glacial lake outburst flood,
Bhutan Himalaya. Mt. Res. Dev. 16 (1), 77–81.
Westoby, M.J., Glasser, N.F., Brasington, J., Hambrey, M.J., Quincey, D.J., Reynolds, J.
M., 2014. Modelling outburst floods from moraine-dammed glacial lakes. Earth Sci.
Rev. 134, 137–159.
Westoby, M., Brasington, J., Glasser, N., Hambrey, M., Reynolds, J., Hassan, M.,
Lowe, A., 2015. Numerical modelling of glacial lake outburst floods using physically
based dam-breach models. Earth Surf. Dyn. 3 (1), 171–199.
Worni, R., Huggel, C., Stoffel, M., 2013. Glacial lakes in the Indian Himalayas—From an
area-wide glacial lake inventory to on-site and modeling based risk assessment of
critical glacial lakes. Sci. Total Environ. 468, S71–S84.
Worni, R., Huggel, C., Clague, J.J., Schaub, Y., Stoffel, M., 2014. Coupling glacial lake
impact, dam breach, and flood processes: a modeling perspective. Geomorphology
224, 161–176.
Yager, E.M., Turowski, J.M., Rickenmann, D., McArdell, B.W., 2012. Sediment supply,
grain protrusion, and bedload transport in mountain streams. Geophys. Res. Lett. 39,
L10402. https://doi.org/10.1029/2012GL051654.
Yamada, T., Sharma, C.K., 1993. Glacier lakes and outburst floods in the Nepal
Himalaya. Int. Assoc. Hydrol. Sci. 218, 319–330.